CT63 only starts from 030 mode

[exxos]

Does all this work properly without the CT60 plugged in ? because if a stock machine works then I would really think that there is something wrong with the CT60. You might be spending time trying to diagnose a fault which does not actually exist.

[foft]

Does all this work properly without the CT60 plugged in ? because if a stock machine works then I would really think that there is something wrong with the CT60. You might be spending time trying to diagnose a fault which does not actually exist.

The machine works fine without the CT60 and also in 030 mode.

I'm using this to learn some more electronics as much as fix it.

Anyway the myth is that the ATX power supplies take too long to start up and thus the 3.3v and 5v lines are not stable before coming out of reset. So people increase C7 to make this work by ... delaying the reset pulse.

Now none of that seems true to me! The 5v and 3.3v lines are stable very quickly and a long time before coming out of reset.

Anyway I'll try the larger C1 now and see if that sorts it.

yellow: 3.3v supply, cyan: 5.0v supply, pink: xreset

ScreenImg(2).png

[foft]

I changed C1 and this does indeed increase the reset pulse length. So there is more time for supplies to stabilize on startup.

I added 100uF so its truly an enormous delay!

Anyway, does it fix the CT63 boot up? Nope.

So next up I guess:
i) I want to understand the 555 circuit more, just because...
ii) I guess I should see what the CT63 is up to. It has a nice expansion connector sitting on top exposing the various address and data lines.

[exxos]

The machine works fine without the CT60 and also in 030 mode.

Then it is unlikely to be a reset issue. I think you are getting hung up on issues which do not actually exist.

Anyway the myth is that the ATX power supplies take too long to start up and thus the 3.3v and 5v lines are not stable before coming out of reset. So people increase C7 to make this work by ... delaying the reset pulse.

Now none of that seems true to me! The 5v and 3.3v lines are stable very quickly and a long time before coming out of reset.

That is a known problem. Some people have had the same with bad caps on the reset circuit when using my power supply. Also poor reset also affects my V1 STE booster.

But as you're working in 030 mode, then this would not appear to be the problem. Indeed if you have checked 3.3V and 5V are reaching levels before reset happens then that is perfectly fine.

I cannot really suggest anything else personally because I have not even used my CT60 since like 20 years ago. I think you said the CT60 does not work in a second Falcon either ? So I really think there is just something inherently wrong with your CT60 IMO. At this point I suggest just returning it and have it tested out.

[foft]

Unfortunately I'm thread hopping again, here and atari-forum (1). Sorry all! :oops:

Anyway this appears to be a known issue which is related to the ramp up of the 3.3v rail vs the 5v rail (2). The hypothesis being that is 3.3v ramps up later then the CPLD is programmed 'too late'.

I'm guessing/wondering if its something to do with putting the 68030 into the right state and if the CPLDs are too late then ...

Anyway what I'm trying to work out if the best way to delay the 5V rail to rise after the 3.3V rail. I'm more a digital electronics person... So, I guess I could have a relay or transistor and when 3.3v is high then let the 5V pass? Or would it be better to have 'power good' drive a couple of transistors and/or relays to enable both 3.3v and 5v?

I was also wondering, there is no public CPLD HDL for these right?

Refs:
1) https://www.atari-forum.com/viewtopic.p ... 39#p438939 (test of my PSU ramp up)
2) https://www.atari-forum.com/viewtopic.p ... 89#p366289 (test of several PSUs ramp up)

[exxos]

But that does not make sense to me.. The whole idea of having a reset signal is to let all voltages stabilised before the machine start actually executing code.

I think you already tested that 5V and 3.3V reaches the correct voltages before reset goes high ?

[foft]

But that does not make sense to me.. The whole idea of having a reset signal is to let all voltages stabilised before the machine start actually executing code.

I think you already tested that 5V and 3.3V reaches the correct voltages before reset goes high ?

You’d think, though… in this case it seems like even with reset low something happens.

In fpga design you don’t need to wire all registers to the reset line. Might be the same with cplds.

Edit:Though in this case it’s the 5v logic that just do something first when in reset?!

[exxos]

You’d think, though… in this case it seems like even with reset low something happens.

In fpga design you don’t need to wire all registers to the reset line. Might be the same with cplds.

Well I think this should be proven first. You could I guess just putting manual toggle switch on the 5V and / or 3.3V line. So you can manually switch one before the other. But don't blame me if you kill the whole thing in doing that :)

[dml]

You’d think, though… in this case it seems like even with reset low something happens.

Which suggests the real issue is one or more devices not observing the reset signal. Figuring out why might yield a simpler solution than trying to order the power up sequence?

[dml]

If someone does want to go down the road of synchronizing the 5v with the 3.3v supply, I would probably look at using a P-MOSFET switch...

5ssj0.png

This will let you use the 3.3v as a logic-level switch, which is isolated from 5V by the gate of the first N-MOSFET. You can use a 2.5v TTL triggered N-MOSFET for this and you can add a diode drop to the gate signal to raise the threshold to 3.0v-3.2v. Or you can add more complexity and use a comparator to set the threshold with a pot.

The P-MOSFET sits on the 5V rail and is 'in front' of the load, so everything is still ground/0v-referenced. This is why two mosfets are needed, because the switching signals and the output both need to be ground-referenced.

It will switch very fast, silent, no sparks, no power loss. Look for a P-MOSFET with low on resistance.

Word of caution - MOSFETS fail into a dead short (and then cook, or explode). So some additional care may be needed to stop the 5V raising up the 3.3V rail in any unfortunate accident...

[EDIT] Actually using the indicated circuit probably does need extra details on the input control - the gate signal is inverted so gate-on means 5v-off. Not difficult to correct but important to get right :)